This revealed a bug in ra_get_spill_benefit where we only considered
the benefit of the first adjacency we were to remove, explaining some
of the ugly spilling I've seen in shaders. Because of the reduced
spilling, it reduces the runtime of glsl-fs-convolution-1 36.9% +/-
0.9% (n=5).
This was recommended in the original paper, but I figued "make it run"
before "make it fast". Now we make it fast. Reduces the runtime of
glsl-fs-convolution-1 by 12.7% +/- 0.6% (n=5).
Our use of the register allocator in i965 is somewhat unusual.
Whereas most architectures would have a smaller set of registers with
fewer register classes and reuse that across compilation, we have 1,
2, and 4-register classes (usually) and a variable number up to 128
registers per compile depending on how many setup parameters and push
constants are present. As a result, when compiling large numbers of
programs (as with glean texCombine going through ff_fragment_shader),
we spent much of our CPU time in computing the q[] array. By keeping
a separate list of what the conflicts are for a particular reg, we
reduce glean texCombine time 17.0% +/- 2.3% (n=5).
We don't expect this optimization to be useful for 915, which will
have a constant register set, but it would be useful if we were switch
to this register allocator for Mesa IR.
Fixes texrect-many regression with ff_fragment_shader -- as we added
refs to the subsequent texcoord scaling paramters, the array got
realloced to a new address while our params[] still pointed at the old
location.
The driver was saying that independend blend functions was not supported,
but it really was. The driver was using the per-target independend blend
factors but the state tracker was only setting the 0th one (per the
Gallium spec).
Fixes a piglit fbo-drawbuffers2-blend regression.
See https://bugs.freedesktop.org/show_bug.cgi?id=33215
The removed semantic check also exists in ast_type_specifier::hir(), which
is a more natural location for it.
The check verified that precision statements are applied only to types
float and int.
* Check that precision qualifiers only appear in language versions 1.00,
1.30, and later.
* Check that precision qualifiers do not apply to bools and structs.
Fixes the following Piglit tests:
* spec/glsl-1.30/precision-qualifiers/precision-bool-01.frag
* spec/glsl-1.30/precision-qualifiers/precision-struct-01.frag
* spec/glsl-1.30/precision-qualifiers/precision-struct-02.frag
Change default value to ast_precision_none, which denotes the absence of
a precision of a qualifier.
Previously, the default value was ast_precision_high. This made it
impossible to detect if a precision qualifier was present or not.
The check is performed only in GLSL versions >= 1.30.
From section 4.3.4 of the GLSL 1.30 spec:
"It is an error to use centroid in in a vertex shader."
Fixes Piglit test
spec/glsl-1.30/compiler/storage-qualifiers/vs-centroid-in-01.vert
The check is performed only in GLSL versions >= 1.30.
Fixes the following Piglit tests:
* spec/glsl-1.30/compiler/interpolation-qualifiers/fs-smooth-02.frag
* spec/glsl-1.30/compiler/interpolation-qualifiers/vs-smooth-01.vert
... and 'centroid varying'. The check is performed only in GLSL
versions >= 1.30.
From page 29 (page 35 of the PDF) of the GLSL 1.30 spec:
"interpolation qualifiers may only precede the qualifiers in, centroid
in, out, or centroid out in a declaration. They do not apply to the
deprecated storage qualifiers varying or centroid varying."
Fixes Piglit test
spec/glsl-1.30/compiler/interpolation-qualifiers/smooth-varying-01.frag.
If an interpolation qualifier is present, then the method returns that
qualifier's string representation. For example, if the noperspective bit
is set, then it returns "noperspective".
This implements the extension by choosing a different set of texture
fetch functions when the texture parameter changes.
Signed-off-by: Dave Airlie <airlied@redhat.com>
If a literal slot isn't used it should be set
to 0 instead of an uninitialized value. Also the
channels for pre R700 trig functions were incorrect.
And most important literals were not counted against ndw,
resulting in an invalid force_add_cf detection.
